AE

A Framework for Analyzing Human Movement in Biomechanics: Goals, Phases, Metrics, and Modeling

Goals and orientation

  • In biomechanics, the overarching aim of movement analysis is to improve human movement in some defined way. This goal orients all subsequent steps (metrics, phases/events, model, instrumentation) and is tied to the specific study goals (e.g., improving performance, reducing injury risk).
  • Decide what you are going to quantify (metrics) based on the goal, the time points you care about, and the available instrumentation.
  • Possible metrics include:
    • Position, velocity, acceleration (kinematic variables)
    • Forces and moments (kinetic variables, obtainable via force plates)
    • Energy-related quantities (e.g., work, power)
  • Instrumentation constraints matter: if a force plate is unavailable, you may be restricted to kinematic data from video; choose metrics that are feasible with available hardware.
  • How metrics are framed matters: you can compute average metrics over a time interval or discrete metrics at a specific time point or at defined events.
  • A model is often created to enable quantitative analysis by simplifying the body into a representation that can be quantified (e.g., rigid body segments).
  • In motion capture, a common approach is to segment the body into rigid bodies and place markers on segments to track their positions and orientations, enabling the computation of joint angles and other segment-to-segment relationships.
  • The four core takeaways to extract from this lecture are:
    • Goals
    • Events and phases
    • Metrics
    • Model and instrumentation
  • After these steps, practical work includes planning data collection protocols, recruiting participants, collecting data, and performing analysis to draw conclusions.
  • Quantitative notes:
    • If there are P phases in a movement, there are E = P + 1 events that delineate the phases. This relationship is a guiding design rule for phase-event structuring. \n - In mathematical terms: \n E = P + 1

Events and phases

  • Break complex movements into clear functional phases (e.g., phase 1, phase 2, …), with time intervals delineated by events that mark transitions between phases.
  • The choice of events is up to the analyst and will influence the granularity of the analysis.
  • Practical example (archery): ready position, draw, anchor reached, aiming, release, follow-through, end. Each event defines a transition to the next phase and can be tied to observable criteria (e.g., a marker’s position crossing a threshold, a velocity sign change, etc.).
  • In hands-on demonstrations (e.g., gymnastics, handspring), time points between events are shown; events may be labeled with numerical values or observed transitions.
  • Event-based delineation can be supported by objective data streams (e.g., kinematic angles, ground reaction forces, etc.) that show clear transitions.
  • For flight-to-landing transitions (or other phases), an event could be when contact with the ground occurs; this marks a transition between phases.

Metrics to quantify movement

  • After defining goals and phases, determine what to quantify (the metrics) to represent performance with respect to those goals.
  • Subgoals: in walking, propulsion or energy generation during the propulsive phase may be a subgoal used to guide metric selection.
  • Metrics can be numerous and may include:
    • Joint angles, angular velocities, angular accelerations
    • Linear positions, velocities, accelerations of markers or segments
    • Planar components: frontal, sagittal, transverse plane motions, depending on the motion of interest
    • Timing metrics, timing of events, durations of phases
  • The number of potential variables can grow quickly, so you must connect metrics to the goals and subgoals to avoid data overload.
  • Practical strategy: leverage coaching cues or domain knowledge to interpret qualitative guidance biomechanically and translate cues into quantitative metrics.
  • This is often an iterative process: initial metrics may not show changes; revise the metric set or the model to better capture the movement of interest.
  • Key reminder: the metric selection and the model should be driven by the research question and feasibility, not by data availability alone.

Modeling and instrumentation

  • You cannot model every detail of the human body; instead, you build a simplified, yet adequate, representation to extract the needed variables (kinematic and kinetic) for analysis and coaching.
  • Common modeling approach: break the body into a small number of rigid segments and quantify their orientation and relative motion to derive joint angles.
  • In optoelectronic motion capture (e.g., Vicon), a standard marker-based model is the Plug-in Gait (PiG) model, which uses a small marker set to define key body segments (e.g., forearm, upper arm, torso, thigh, shank) and compute segment positions and orientations.
  • This model supports gait analysis and other movements by providing a framework to compute joint angles and segmental motions.
  • For movements outside well-studied domains, you may need to design a custom model (marker set or segment definitions) tailored to the movement and questions of interest.
  • Marker-based tracking allows extraction of defined metrics over time; other instrumentation options include:
    • Force transducers/plates (for kinetic data and ground reaction forces)
    • Motion capture (optical, marker-based)
    • Inertial measurement units (IMUs)
    • Eye tracking or other sensor modalities as relevant to the task
  • A key challenge is deciding what instrumentation is necessary to quantify the chosen metrics effectively and efficiently.
  • Iterative design of the model and instrumentation is common: update markers, adjust sensor types, or redefine how segments are represented based on what provides the most robust and useful data for the questions at hand.
  • The goal of modeling and instrumentation is to provide a deliverable set of variables that can be used to compare performances, coaching cues, and training outcomes in a repeatable, objective way.

Case study: archery (example of goals, phases, and metrics)

  • Goals you might consider for archery include:
    • Target accuracy (hitting the target)
    • Projectile distance (horizontal range) or projectile momentum upon impact (relevant for hunting or distance optimization)
    • Injury avoidance and tissue sustainability for repeated attempts (long-term safety)
    • Higher-order goals like recreational enjoyment, strength development, or winning competitions
  • Subgoals often relate to technique and movement quality that support the higher-order goals (e.g., stability, efficiency, repeatability).
  • Event-based breakdown for archery (one possible design):
    • Ready position (initiation of movement) → Draw phase (pulling the arrow back) → Anchor position reached (a defined pose) → Aiming phase (steady hold/visual alignment) → Release point (arrow leaves the hand) → Follow-through (post-release motion) → End of movement (hand comes to rest)
  • Alternative design: focus on the arrow’s flight phase instead of the archer’s fixation on follow-through; the final event could be the arrow contacting the target.
  • This demonstrates a divergence in design decisions: you can emphasize the performer (the archer) or the product (the arrow flight) depending on the research focus.
  • Instrumentation considerations for archery include:
    • Detecting the moment of release (e.g., markers on the hand or arrow, velocity or acceleration signatures, or pressure sensors)
    • Identifying phase transitions (e.g., start of draw, anchor, release) using kinematic signals or force data
    • Deciding whether to track the archer’s movement (joint kinematics) or the arrow’s flight (projectile tracking) or both
  • Metrics in archery could include: joint positions/velocities, hand/wrist angles, draw length dynamics, timing of release, timing of each phase, stability measures during aiming, and the speed/angle of the follow-through.
  • Planning for measurement requires choosing which joints to monitor (e.g., wrist, elbow, shoulder) and in which planes of motion to focus on, recognizing that increasing the number of tracked joints expands the metric space quickly.

Practical considerations and iteration

  • This analysis is iterative: start with a plausible metric set, collect data, examine whether the metrics capture meaningful changes, and adjust accordingly.
  • If a movement improves in performance but the chosen metrics do not reflect it, reconsider either the metrics or the underlying model.
  • Coaching cues can guide metric selection (e.g., if a cue emphasizes alignment, quantify alignment angles or related kinematic variables).
  • Real-world data collection involves deciding on a practical protocol, recruiting participants, and ensuring data quality and consistency.
  • Instrumentation choices should balance practicality, cost, and information content: markers sets, force plates, IMUs, EMG, and possibly other sensors depending on the question.
  • The overarching aim is to provide a framework that allows objective comparison across attempts, rehabilitation progress, or optimization of performance, bridging clinical rehabilitation and performance optimization contexts.

Time-series and objective phase delineation example

  • A time-series example used to delineate phases is elbow angle displacement during a throwing movement: one can observe clear changes in the direction of elbow displacement (e.g., from negative to positive displacement) that indicate a transition from eccentric (lengthening) to concentric (shortening) muscle action.
  • This illustrates how objective data streams (angles over time) can identify phase transitions without subjective judgments.
  • Such objective markers can be used to define phase boundaries in data analysis and to validate the chosen events.

Model complexity and practical design choices

  • The level of detail in a model should be sufficient to answer the research question but not so complex that it becomes infeasible or introduces unnecessary noise.
  • Established models (e.g., Plug-in Gait) provide a starting point for many common movements (gait, basic limb motions) and can be adapted for new tasks.
  • When dealing with novel movements, create a bespoke model (marker set and segment definitions) that captures the essential mechanics relevant to your questions.
  • The marker-based approach relies on a manageable number of markers to estimate segment positions and orientations over time, enabling computation of joint angles and other metrics.
  • Practical guidance for students: leverage existing literature and established models to anchor your approach, then iterate to tailor the model to your specific movement and questions.

Reconciling goals, phases, metrics, and models: a practical workflow

  • Start with the high-level goal: what improvement or outcome do you want to achieve?
  • Define the functional phases and the events that delineate the transitions between phases; remember the E = P + 1 rule.
  • Determine the metrics that best represent progress toward the goal, including subgoals if useful.
  • Design or select a model and instrumentation capable of producing those metrics with sufficient reliability.
  • Plan data collection (protocol, participant recruitment) and data processing methods.
  • Iterate: assess whether the metrics capture movement performance as intended; refine the model, metrics, or instrumentation as needed.
  • The aim is to enable objective comparison across movements, guide coaching or rehabilitation decisions, and support optimization of performance over time.

Final takeaway

  • The four core pillars to focus on are: goals, events/phases, metrics, and model/instrumentation. Everything else (data collection, recruitment, analysis, conclusions) follows from these foundations.
  • This framework helps translate high-level questions about movement into actionable, quantitative analyses that can inform coaching, rehabilitation, or performance optimization.
  • The approach emphasizes objective delineation of phases, careful metric selection aligned with goals, and a thoughtful model that balances simplicity with the ability to answer the research questions.